more code review

Author: petrelharp

python/tests/test_metadata.py

@@ -649,8 +649,49 @@ def test_disallow_duplicate_keys(self):
         ):
             tskit.MetadataSchema(schema)
 
-    def schema_with_binary(self, num_blobs):
-        # produces a json+struct schema having some number of integers
+    def test_round_trip_with_struct_and_json(self):
+        schema = {
+            "codec": "json+struct",
+            "json": {
+                "type": "object",
+                "properties": {
+                    "label": {"type": "string"},
+                    "count": {"type": "number"},
+                    "stuff": {
+                        "type": "array",
+                        "arrayLengthFormat": "h",
+                        "items": {"type": "number"},
+                    },
+                },
+                "required": ["label"],
+            },
+            "struct": {
+                "type": "object",
+                "properties": {
+                    "xyz": {
+                        "type": "array",
+                        "arrayLengthFormat": "H",
+                        "items": {"type": "number", "binaryFormat": "i"},
+                    },
+                },
+            },
+        }
+        ms = tskit.MetadataSchema(schema)
+        for v in [[], [0, 2, 12], [5] * 1000]:
+            row = {
+                "label": "abcdef xyz",
+                "count": 7,
+                "b0": 123,
+                "b1": 0,
+                "stuff": [1, 3, 2, -1.5],
+                "xyz": v,
+            }
+            encoded = ms.validate_and_encode_row(row)
+            out = ms.decode_row(encoded)
+            assert out == row
+
+    def schema_with_binary(self, num_binary_ints):
+        # produces a json+struct schema having num_binary_ints integers
         # encoded in binary, labeled b0, ... bX
         schema = {
             "codec": "json+struct",
@@ -667,28 +708,13 @@ def schema_with_binary(self, num_blobs):
                 "properties": {},
             },
         }
-        for j in range(num_blobs):
+        for j in range(num_binary_ints):
             schema["struct"]["properties"][f"b{j}"] = {
                 "type": "integer",
                 "binaryFormat": "i",
             }
         return tskit.MetadataSchema(schema)
 
-    def test_round_trip_with_struct_and_json(self):
-        schema = dict(self.schema_with_binary(2).schema)
-        # adds a 'xyz' property to the schema, which stores an array of numbers
-        schema["struct"]["properties"]["xyz"] = {
-            "type": "array",
-            "arrayLengthFormat": "H",
-            "items": {"type": "number", "binaryFormat": "i"},
-        }
-        ms = tskit.MetadataSchema(schema)
-        for v in [[], [0, 2, 12], [5] * 1000]:
-            row = {"label": "abcdef xyz", "count": 7, "b0": 123, "b1": 0, "xyz": v}
-            encoded = ms.validate_and_encode_row(row)
-            out = ms.decode_row(encoded)
-            assert out == row
-
     @pytest.mark.parametrize("k", (0, 1, 5, 1001))
     def test_byte_alignment(self, k):
         # We want to test whether the binary portion begins byte-aligned.
@@ -704,8 +730,14 @@ def test_byte_alignment(self, k):
         bytes_per_blob = len(struct.pack("i", 0))
         for s in [
             "",
+            "a",
+            "ab",
             "abc",
-            "superfragilisticexpialodocious",
+            "abcd",
+            "abcde",
+            "abcdef",
+            "abcdefg",
+            "abcdefgh",
             " " * 1000 + "foo" + " " * 1000,
         ]:
             row = {"label": s, "count": 7}